Devices that interact with users primarily through audio commands given by the user, and audio feedback provided to the user through a built-in speaker have recently become more popular. With such devices, users can ask questions, dictate messages, listen to written messages that are read out loud by the device, play music, and perform other tasks through audio interaction without having to interact with a graphical user interface. Some users may wish to connect such devices to external speakers, but this may limit the usefulness of the device if the external speakers are off, or otherwise not playing the audio output by the device.
Accordingly, new methods, systems and media for controlling audio output are provided.